A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 12, 14:45 UTC
[BugFix] Use unique ids for different transcription prompts (#29372)
Signed-off-by: Nick Hill <nhill@redhat.com>
Nick Hill committed
Nov 24, 2025 at 22:55 UTC
7992324f23478bebf5e39542a4ce198cd7a1ab2a
1 file changed
+2
-2
vllm/entrypoints/openai/speech_to_text.py
+2
-2
@@ -201,10 +201,10 @@ class OpenAISpeechToText(OpenAIServing):
201
self.engine_client.generate(
202
prompt,
203
sampling_params,
204
- request_id,
204
+ f"{request_id}_{i}",
205
lora_request=lora_request,
206
)
207
- for prompt in prompts
207
+ for i, prompt in enumerate(prompts)
208
]
209
except ValueError as e:
210
# TODO: Use a vllm-specific Validation Error